Answers are grounded in this paper and cite their sources.
Longitudinal analysis reveals factors influencing support behaviors in an online diabetes community, suggesting targeted interventions.
Introduction and Objective: Online health communities (OHCs) provide essential social support for individuals with type 1 diabetes (T1D), yet how support-seeking behaviors evolve into support-providing roles over time remains unclear. This study examined temporal patterns and factors associated with this transition in the largest T1D online community in China. Methods: We conducted a longitudinal analysis of 378,304 posts from 7,803 users (2019-2024). Posts were classified into information and emotional support-seeking and -providing behaviors using BERT-Base-Chinese, and emotional expressions were identified with Chinese Clinical BERT. Behavioral state transitions (baseline, seeker, provider) were modeled using hidden Markov models (HMM) with covariates including disease duration and community engagement, based on 6-month time windows. Individual trajectories were reconstructed using the Viterbi algorithm and clustered via dynamic time warping (DTW). Results: Overall, 46.2% of users experienced at least one behavioral state transition. Transitions from support seeking to support providing (28.4%) occurred more frequently than reverse transitions (9.7%). Longer disease duration (β=0.42, p<0.01) and longer community engagement (β=0.31, p<0.05) were associated with a higher likelihood of becoming a support provider. Higher emotional intensity was also positively associated with this transition (β=0.27, p<0.05). Three distinct behavioral trajectories were identified: emotion-driven (32%, rapid transition), rational-accumulative (45%, gradual transition), and fluctuating (23%, unstable). Conclusion: Users in T1D online health communities show clear longitudinal shifts from support seeking to support providing, associated with greater disease experience and sustained community engagement. These findings highlight opportunities for targeted digital interventions in chronic disease management.
No takes yet. Share an insight, caveat, or question.
ZHU et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: